Cross-kingdom signalling regulates spore germination in the moss Physcomitrella patens

Plants live in close association with microorganisms that can have beneficial or detrimental effects. The activity of bacteria in association with flowering plants has been extensively analysed. Bacteria use quorum-sensing as a way of monitoring their population density and interacting with their environment. A key group of quorum sensing molecules in Gram-negative bacteria are the N-acylhomoserine lactones (AHLs), which are known to affect the growth and development of both flowering plants, including crops, and marine algae. Thus, AHLs have potentially important roles in agriculture and aquaculture. Nothing is known about the effects of AHLs on the earliest-diverging land plants, thus the evolution of AHL-mediated bacterial-plant/algal interactions is unknown. In this paper, we show that AHLs can affect spore germination in a representative of the earliest plants on land, the Bryophyte moss Physcomitrella patens. Furthermore, we demonstrate that sporophytes of some wild isolates of Physcomitrella patens are associated with AHL-producing bacteria.

Plants do not exist in isolation in the environment, but interact with a wide array of organisms from all kingdoms including bacteria, fungi, animals and other plants. These interactions can have profound effects on plant fitness, growth and development. In addition to pathogenicity or parasitism, interactions between plants and other organisms can be beneficial. Examples include interactions between fungi and plants in the form of mycorrhizae and interactions between plants and bacteria [1][2][3] .
Interactions between plants and microorganisms have become more elaborate during plant evolution. Mycorrhizal interactions are beneficial in liverworts, one of the earliest diverging groups of land plants, where the association between liverworts and fungi boosts plant photosynthesis, growth, fitness and nitrogen/phosphorus uptake 4,5 . In later-diverging plants including flowering plants, interactions with microorganisms have increased in complexity. For example, in legumes (Fabaceae, including beans and pulses) plant root cells are surrounded by a group of proteobacteria (Rhizobia) and form a symbiotic root nodule 6 while actinobacteria from the Frankia genus can form nodules with a wide range of plant families 7,8 . The bacteria gain carbon from the plant, while the plant gains nitrogen from the bacteria. Not all plant-bacterial interactions are so highly specialised: many bacteria in the rhizosphere contribute to plant productivity and gain from plants in return 9 .
It has become clear that flowering plants can respond to bacterial signalling molecules that alter plant growth and development, representing inter-kingdom interaction 10,11 . The perception of bacteria by plants is of significant importance in terms of monitoring their surroundings and thus being able to respond accordingly to enhance their chances of survival 12 .
A key way in which bacteria communicate with one another is via diffusible quorum sensing (QS) molecules that are used to monitor and respond to population density within a colony or biofilm [13][14][15][16][17][18][19] . One well-characterised subset of QS molecules that affect plants' behaviour are the N-acylhomoserine lactones (AHLs) 11,12,[20][21][22][23][24] . AHLs are We therefore hypothesised that AHLs might affect development in early-diverging land plants. In this paper, we show that synthetic AHLs can promote spore germination in the model moss species Physcomitrella patens 77,78 in a lab-based assay. Moreover, sporophytes from wild isolates of Physcomitrella patens are associated with AHL-producing bacteria, suggesting these bacteria may influence spore germination in the environment through the production of AHL signal molecules.

Results
AHLs promote Physcomitrella spore germination at sub-micromolar concentrations but inhibit spore germination at concentrations above 1 µM. Previous studies have shown that AHLs at concentrations of 1-10 µM can promote root growth in the model flowering plant Arabidopsis 22,45,47,48 . In algae, AHLs at 2-10 µM can promote spore release 73 or reduce the progress of growth and development from spores 72 . We tested the effect of a range of AHLs with different carbon N-acyl chains lengths at 0.1 µM and 1 µM concentrations on the spore germination of the model moss Physcomitrella patens. All AHLs (C4-HSL to C12-HSL) induce a significantly faster spore germination rate compared to a solvent-only control (Fig. 1). C4-HSL and C6-HSL show a similar promotion of germination at both 0.1 µM and 1 µM concentrations (Fig. 1A,B). C8-HSL appears to have slightly more germination-promoting activity than the shorter chain AHLs and is more potent at 1 µM than 0.1 µM (Fig. 1C). C10-HSL and C12-HSL are the most potent germination-promoting AHLs, being more effective at 0.1 µM than at 1 µM concentration (Fig. 1D,E).
In Arabidopsis, 50-100 µM concentrations of AHLs inhibit root growth 21,45,48 while in the seaweed Ulva spore germination and early development is reduced with just 5 µM AHLs 72 . We tested the effects of a range of AHLs at 5 µM on Physcomitrella spore germination and found that AHLs could inhibit spore germination ( Fig. 2A). The effect appeared strongest with C10-HSL, which also inhibited germination at 10 µM, in a dose-dependent manner (Fig. 2B). Taken together, these data show that AHLs, particularly those with longer chain length, accelerate spore germination when at low (≤1 µM) concentrations and inhibit spore germination at higher (5-10 µM) concentrations.
Chain length and side group substitution affect the activity of AHLs against Physcomitrella spore germination. To investigate whether changing the side group of the AHL had an effect on biological activity, we assayed the spore-germination-promoting activity of C4-C12 AHLs, namely the N-acyl version (as before) and also the 3-oxo (3-O) and 3-hydroxy (3-OH) substituted forms. Our "snapshot" data ( Fig. 3A) indicated potential differences in potency between the different side chains, particularly for AHLs with longer carbon chain. To investigate these differences further, we assayed spore germination in the presence of 3-OH and 3-O substitutions of the C10 and C12 HSLs, which consistently through this study showed some of highest activity, over a range of concentrations from 2 nM to 1 µM (summarised in Fig. 3B; data in Supplemental Fig. 1 and 2). The N-acyl variants of C10-and C12-HSL showed the greatest spore germination-promoting activity at 2-10 nM ( ). These data demonstrate that both chain length and side group substitution can affect the biological activity of exogeneously-applied synthetic HSLs on Physcomitrella spore germination. As seen in 79 there is variability in control germination rate between different spore batches (e.g. Fig. 1 versus Fig. 3) but this does not affect the trends seen with different treatments.
Sporophytes from wild isolates of Physcomitrella are associated with multiple species of bacteria, some of which produce AHLs. To determine whether the observed effects of synthetic AHLs on Physcomitrella patens spore germination in a lab-based assay might have relevance to wild populations found in the environment, sporulating Physcomitrella plants were firstly collected from 3 different locations in the UK with a view to determine the bacterial populations associated with them and the ability of these bacteria to produce AHLs. A total of 12 plants were selected from each sampled location and the sporophyte from each plant was isolated to enable isolation of its associated bacteria (a consortium from each sporophyte). Bacterial consortia were obtained from 30 out of 36 sporophytes. Each individual sporophyte's (assumed mixed) bacterial populations were taken through multiple rounds of streaking to isolate individual strains associated with Physcomitrella. To identify each strain, a fragment of the 16S rRNA gene was amplified from genomic DNA and sequenced. Our sampling of bacteria associated with Physcomitrella sporophytes identified largely Proteobacteria from the class Gamma-proteobacteria. Bacteria of the genus Pseudomonas were found at all 3 sites (at least 5 different species), as was Stenotrophomonas (2 species). Serratia (2 species) were isolated from 2 sites and Acinetobacter, Aeromonas and Rahnella were each recovered from a single site. The gram-positive bacteria Microbacterium (Actinobacteria) and Bacillus (Firmicutes) were each found at a single site (Table 1).
For an initial survey of whether the isolated bacterial consortia from each sporophyte could produce AHLs, consortia were analysed by a mass-spectrometry (LC-MS/MS) approach, which demonstrated that consortia from all three locations could produce AHLs although these were detected to only a limited extent from the Lindley site (Table 2).
To determine whether the individual bacteria isolated from wild Physcomitrella could produce detectable AHLs, cultures of the Gram-negative bacterial isolates were subjected to AHL analysis by LC-MS/MS. Just under half of the bacterial isolates from each of Chew Valley and Stocks reservoir produced detectable AHLs, while no AHLs were detected from the bacteria from Lindley. Representatives of N-acyl, 3-O and 3-OH from C4 to C10 chain length were detected (Fig. 4). Overall, the most frequently detected AHLs were C6-HSL and 3-O-C8-HSL. patens spores. Z-tests indicated significant differences in germination between treated and untreated spores on days 3 and 11 (P > |t| 0.0002). (B) C6-HSL promotes germination of P. patens spores. Z tests indicated significant differences in germination between treated and untreated spores on days 3 and 11 (P > |t| 0.0002). Treatment with the lower concentration of 0.1 μM was more effective in promoting germination on days 3 and 11 when compared to 1 μM. (C) C8-HSL promotes germination of P. patens spores. Z tests indicated significant differences in germination % between treated and untreated spores on days 3, 7 and 11 (P > |t| 0.0002). (D) C10-HSL promotes P. patens spore germination. Z tests indicated significant differences in germination % between treated and untreated spores on days 3, 7 and 11 (P > |t| 0.0002). Treatment with the lower concentration of 0.1 μM was significantly more effective in promoting germination on days 3 and 7 when compared to 1 μM. (E) C12-HSL promotes P. patens spore germination. Z tests indicated significant differences in germination % between treated and untreated spores on days 3, 7 and 11 (P > |t| 0.0002). Treatment with the lower concentration of 0.1 μM was significantly more effective in promoting germination on days 3 and 7 when compared to 1 μM. In all experiments, final germination efficiency was not affected with all treatments achieving a final germination of over 95%. Representative of more than 5 biological repeats. Error bars represent ± SEM. www.nature.com/scientificreports www.nature.com/scientificreports/ The most frequently detected AHL in bacteria from Chew Valley was 3-OH-C10-HSL, whereas that from Stocks was 3-O-C6-HSL.
Taken together, these data show that some of the bacteria associated with Physcomitrella sporophytes from different geographical locations can produce a range of AHLs.

Discussion
Our experiments show for the first time that synthetic AHLs can affect the spore germination of an early diverging land plant, the bryophyte Physcomitrella patens, in a lab-based assay. Low (<1 µM) concentrations of AHLs promote spore germination whilst higher concentrations (5-10 µM) inhibit spore germination. In general, AHLs with longer chain length (C8-C12) have a more potent effect than C4-C6 AHLs and side-group substitutions change the potency of germination-promoting activity with 3-O and 3-OH substitutions generally showing a slight reduction in potency.
The inhibitory effect of higher concentrations of AHLs is reminiscent of their effect in the green seaweed Ulva where >5 µM AHLs can inhibit the early development and growth of new plants from zoospores 72 . Higher concentrations (25-125 µM) of long-chain AHLs can also inhibit Ulva spore swimming speed to promote settlement with 3-O substitutions showing the greatest inhibition 74 . The effect of sub-micromolar concentrations AHLs was not investigated in these experiments.
In land plants, the effect of AHLs on germination of the desiccation-resistant dispersal units, namely spores (in Bryophytes, Lycophytes and ferns) or seeds (in Gymnosperms and Angiosperms) is not well studied. So far, a single study shows that priming of winter wheat (Triticum aestivum L.) seeds with C6-HSL (~9 ng AHL per seed) improves their germination and subsequent growth, development and biomass production 80 . Thus, a potential role in germination control for AHLs is present across plant-and algal taxa, although whether this is as a result of convergent or divergent evolution is unknown.  22,45,47,48,51 while >10 µM of C10-C14 AHLs inhibit primary root growth in Arabidopsis seedlings 21,48 . Moreover C10 and C12 AHLs promote root branching and increases root hair formation at 12-96 µM 21 . Inhibitory effects of AHLs on the Arabidopsis root involve changes in cell division and differentiation 21 . This biphasic pattern (growth stimulation of Arabidopsis primary root at low concentrations, growth inhibition at higher concentrations) is reminiscent of what we see with Physcomitrella spore germination (Figs. 1-3). In general, longer-chain AHLs (C10, C12) have more potent effects, as we saw with Physcomitrella spore germination in this paper, although the concentrations required for an effect in Arabidopsis are higher (≥1 µM) than in Physcomitrella (2nM-1 µM). In barley, 10 µM C6-HSL promotes seedling growth 35 .
Several studies hint at the molecular mechanisms underlying the effects of AHLs on Arabidopsis root growth. Transcriptome-and qRT-PCR approaches coupled with mutant studies implicate several transcription factors in the response, including AtMYB44 51 , in addition to G-protein signalling 47,48 and calmodulin/calcium signalling 45,46 . Interestingly, a role for changes in intracellular calcium signalling has also been implicated in Ulva spore settlement, which is also affected by AHLs 81 .
There is considerable overlap between the bacteria we isolated from Physcomitrella sporophytes and the bacteria found in assocated with the peat moss Sphagnum 59 in which Pseudomonas, Rahnella, Serratia, Stenotrophomonas and Microbacterium are all present but Aeromonas and Acinetobacter were not detected. No Bacillus was detected in Sphagnum, although Paenibacillus (Firmicutes) was, along with additional Beta-proteobacteria, Bacteroidetes, and Actinobacteria 59 .
The bacteria isolated from mosses are generally different from those isolated from Ulva: predominantly Alpha-proteobacteria and Bacteroidetes, although Microbacterium has been isolated from all three species [82][83][84] . Most of the genera of Gamma-proteobacteria isolated (Pseudomonas, Serratia, Acinetobacter, Aeromonas) are AHL producers 26,[85][86][87][88] . However, it is important to note that only a small fraction (<1%) of all bacteria that exist in a particular environment can be grown in the lab on standard growth media 89 so there may be many other AHL-producing bacteria associated with Physcomitrella in the wild.
Single-species AHL analysis showed that, as expected, many of the Pseudomonas isolates, most of the Serratia isolates and one of the Aeromonas isolates produced AHLs (Fig. 4). Moreover, an isolate of Stenotrophomonas from each of the Chew Valley and Stocks reservoir sites also produced AHLs: this genus has not previously been -C12 chain length was tested against solvent control for effects on spore germination. A "snapshot" of germination on day 3 is shown. Asterisks represent significant (*p < 0.05; **p < 0.01) differences between a treatment and solvent control using a Kruskal-Wallis test and a Dunn's post-hoc test. Generally, longer chain AHLs stimulate germination more, and AHLs without or with 3-O substitutions appear more potent than those with 3-OH substitutions at this concentration. (B) Summary of the optimal concentrations of AHLs for promoting Physcomitrella spore germination: full data is shown in Supplemental Fig. 1.  (2020)  www.nature.com/scientificreports www.nature.com/scientificreports/ found to make AHLs as it normally makes DSF-type quorum sensing molecules 90 which also induce growth promoting traits on plants 91 . Two Pseudomonas fluorescens isolates from the Stocks reservoir showed AHL production even though, to our knowledge, no strains from this species have been reported before to produce these QS molecules. Unexpectedly, none of the individual isolates from Lindley produced AHLs (Fig. 4) despite several attempts, suggesting that they have either lost the ability to produce these molecules or, under the in vitro growth conditions used, they only make AHLs below the lower limit of detection for the LC-MS/MS system used.
In summary, we have characterised for the first time the effect of bacterial quorum sensing molecules, AHLs, on the development of a non-flowering land plant, the moss Physcomitrella patens. AHLs promote Physcomitrella spore germination at sub-micromolar concentrations, but inhibit germination at higher concentrations, in a biphasic pattern reminiscent of the AHL effect on root growth in flowering plants. We have shown that a range of bacteria, some of which produce AHLs, are associated with Physcomitrella sporophytes isolated from the wild. Future research could include a metagenomic analysis to identify all bacteria (including those that are uncultivable) associated with Physcomitrella, work with mutant strains of bacteria deficient in AHL production, analysis of calcium signalling in moss spores upon AHL application, isolation of moss mutant strains that cannot respond to AHLs, or transcriptomic/proteomic analysis of moss spores treated with AHLs.

Materials and Methods
Moss spore germination assays. Germination assays were carried out as in 92 . Briefly, spores from at least 3 age-matched sporophytes were used within each assay with three sporophytes' worth of spores used for every 10 Petri dishes (9 cm diameter). Sporophytes were bleached in groups of two to three in 1 ml 25% Parozone TM (Jeyes Group, Thetford, UK) for 10 min and then washed three times in 1 ml sterile distilled water (10 min each) in a sterile flow cabinet. The sporophytes were then crushed in 100-200 μl of sterile water to release the spores. Spores were diluted down in sufficient sterile distilled water to allow plating of 500 μl of spore solution per Petri dish. Spores were plated on cellophane-overlaid BCD moss growth medium (   Isolation and purification of bacteria. Initially, 12 sporophytes from each location were placed on individual Luria broth (LB)-agar plates and bacteria were allowed to grow out from the sporophyte for 2 days at 28 °C in the dark (lower temperatures favoured growth of fungal contamination). The majority of sporophytes were associated with bacteria that could be grown on LB-agar, giving rise to bacterial consortia. These consortia were further purified by taking them through 3-4 rounds of streaking (giving rise to multiple single colonies) as appropriate, growing on LB-agar at 28 °C overnight to obtain multiple pure bacterial isolates (identifiable by morphology and colour) for gDNA isolation, sequence identification and AHL detection. Stock plates for each strain were generated from a single colony and colonies from these plates were inoculated into liquid culture to make permanent glycerol stocks in 25% glycerol, 75% LB. Numerical values on the legend scale are of peak area for detected analytes. A positive detection of an AHL was considered as a chromatographic peak that has a signal to noise ratio of at least 5, displaying a peak retention time that matched that of authentic AHL synthetic standards. www.nature.com/scientificreports www.nature.com/scientificreports/ Bacterial identification. Bacterial isolates were identified to genus-, or where possible species-level.
Bacterial cultures were grown in LB from single colonies and genomic DNA was extracted using a Qiagen Blood and Tissue DNeasy kit (Qiagen, Hilden, Germany) following the manufacturer's instructions. Partial 16S rDNA fragments (~2 kb) were amplified from 10-30 ng of genomic DNA by PCR using the forward primer 27 F (AGA GTT TGA TCC TGG CTC AG) and reverse primer 1522 R (AAG GAG GTG ATC CAG CCG CA). PCR was carried out using Velocity proofreading DNA polymerase (Bioline) according to manufacturer's instructions. The PCR cycling conditions were a denaturation of 94 °C for 2 min followed by 30 cycles of 94 °C for 30 sec, 58 °C for 30 sec and 72 °C for 1 min, then a final extension of 72 °C for 5 min. PCR products were purified using a GeneJET PCR purification kit (Thermo Fisher) and were sequenced using both the forward and reverse primers via capillary sequencing on an ABI3730 machine (Applied BioSystems). Raw sequence reads viewed in SnapGene version 1.4 and were trimmed and refined by eye from the peak trace where necessary. Where possible, forward and reverse sequences were aligned and combined to generate a single consensus sequence. Sequences were analysed by BLASTN 95  For the analysis by LC-MS/MS, chromatography was achieved using a Shimadzu series 10AD LC system. The LC column, maintained at 40 °C, was a Phenomenex Gemini C18 (3.0 µm, 100 × 3.0 mm). Mobile phases A and B were 0.1% (v/v) formic acid in water and methanol respectively. The flow rate throughout the chromatographic separation was 450 µL/min. The binary gradient initially began at 10% B for 1.0 min, increased linearly to 50% B over 0.5 min, then to 99% B over 4.0 min. This composition remained for 1.5 min, decreased to 10% B over 0.1 min, and stayed at this composition for a 2.9 min period of re-equilibration.
For the MS detection of eluting AHLs, an Applied Biosystems Qtrap 4000 hybrid triple-quadrupole linear ion trap mass spectrometer equipped with an electrospray ionisation (ESI) interface was used. Analysis was conducted with the MS operating in positive electrospray (+ES) multiple reaction monitoring (MRM) mode, screening the LC eluent for specific unsubstituted, 3-O and 3-OH AHLs with even numbered acyl chain length from 4-14 carbons long, and the deuterated internal standard, comparing the retention time of detected analytes with authentic synthetic standards.

Data availability
All sequence data generated in this study has been deposited at GenBank and the accession numbers are given in Table 1. The raw data supporting the conclusions of the germination assays and AHL quantification will be made available by the authors, without undue reservation, to any qualified researcher.